Internal compiler bug using an array of integer(kind=1)

[Steve Kargl]

On Sat, Oct 11, 2008 at 09:44:54AM +0200, Paul Richard Thomas wrote:
> Hi Steve,
> 
> 
> I'll commit it this afternoon - thanks for the patch.
> 
> >                === gfortran Summary ===
> >
> > # of expected passes            28266
> > # of unexpected successes       23
> 
> What are these?
> 

Some of the default_format_denormal_?.f90 pass for
REAL(10).  If I change the xfail to *-*-freebsd[4-7],
the unexpected successes turn into expected passes,
but I then get new unexpected FAILs.
 
-- 
Steve